276 



Lord Rayleigh. On the Intensity of [Nov. 18, 



from the observed loss of light transmitted. The transmitted light 

 is about 91 per cent. ; and thus an error arising from the neglect of 

 absorption, or from imperfect matches, amounting, say, to 1 per cent., 

 leads to a relative error of more than 10 per cent, in the estimated 

 reflection. The importance of this consideration may be illustrated 

 by Professor Rood's actual results. In the first case recorded by him 

 the observed transmission was 91*440 as against the theoretical 

 91* 736. The difference 0'296 is indeed very small reckoned upon the 

 transmitted light ; but if we translate the results into terms of the 

 reflected light, they present a different appearance. On the sup- 

 position that the whole loss in the transmitted light is due to 

 reflection, we get for the intensity of the reflected light 8 '560, which 

 is to be compared with the theoretical 8'264. The difference 0*296 is 

 now some 3J per cent., and is thus by no means insignificant. In the 

 other case given by Professor Rood the discrepancy is greater still, 

 amounting to 7 per cent. It may be remarked that in both cases 

 the amount of the reflection appears to be in excess of that given by 

 Young's formula. But the cause may lie in the assumption that the 

 whole failure of transmission is due to reflection. And whatever the 

 explanation may be, we can hardly agree with Professor Rood when 

 he concludes that these experiments show " that the reflecting power 

 of glass with the above index of refraction, conforms in the closest 

 manner to the predictions of theory" 



In the hope of being able to deal directly with the reflected light, 

 I made a great many trials of various devices during the spring of 

 1885, but without finding anything satisfactory. Indeed, at one time, 

 I had almost come round to the opinion that the difficulties of 

 measuring the reflected light were so great that Professor Rood had 

 shown a wise discretion in declining to face them, and that after all 

 the best results would perhaps be reached through measurements of 

 the transmitted light checked by the use of plates of different thick- 

 nesses so as to eliminate absorption.* If, indeed, we give up the 

 perpendicular incidence, the objection founded upon the relatively 

 small quantity of light reflected may be met ; for at an incidence of 

 70°, about half the light (polarised in the plane of incidence) is 

 reflected. In such an experiment it would of course be necessary to 

 determine accurately the angle of incidence. 



The difficulties referred to have their origin in the necessary 

 alteration in the course of the light by the act of reflection. The 

 direct and reflected light cannot be interchanged in any simple 

 manner, and the shift necessary to bring about the substitution may 

 easily lead to systematic error. In the apparatus (presently to be 

 described) to which I was finally led, the difficulty seems to be fairly 



* I was not aware until lately that Sir John Conroy was at work in this 

 direction. 



